The yeast mitochondrial porin represses Snf1/AMP kinase signaling to attenuate viral replication

Abstract

Although fungi are broadly infected with mycoviruses, the antiviral mechanisms fungal cells use to oppose viral replication are not well understood. Here, we discover a new mitochondrially controlled signaling mechanism in the budding yeast Saccharomyces cerevisiae that limits replication of L-A, an RNA mycovirus that endemically infects this organism. We show that Por1, the mitochondrial voltage dependent anion channel, prevents hyper-replication of L-A in stationary phase cells that have exhausted media nutrients. By investigating known stationary phase regulators, we find that deletion of the AMP-activated kinase homolog SNF1 reverses hyper-replication of L-A observed in por1Δ cells. This epistatic relationship suggests that Por1 negatively regulates Snf1 in stationary phase cells and derepressed Snf1 promotes L-A hyper-replication. We confirm this model, first demonstrating that POR1 prevents the accumulation of activated Snf1 throughout stationary phase. By investigating Snf1 signaling targets, we show that this POR1-SNF1 regulatory mechanism acts in stationary phase cells to limit amino acid availability that sustain L-A replication. POR1-SNF1 signaling represents a novel physiological control mechanism to limit viral replication in a eukaryotic cell.

Introduction

Recent sequencing studies have revealed a vast array of RNA viruses that infect many fungal species (Sato and Suzuki 2023). Fungal mycoviruses are well known to persist as endemic infections transmitted through cell division or fusion, with no known extracellular route. Although sometimes believed to be asymptomatic, mycoviruses can profoundly affect their fungal hosts and are better understood of as spanning the symbiotic spectrum as opposed to being inconsequential travelers (Wu and Li 2025). For example, mycovirus infections of the rapeseed phytopathogen Leptosphaeria biglobosa promote survival of the fungus through seasonal periods of high temperature that intervene crop cycles (Zhou et al. 2026). In other cases, mycovirus infection apparently disadvantages the host as seen in phytopathogens where “hypoviruses” limit the fungus's ability to mount robust plant infections (Kotta-Loizou 2021). Recent findings show that mycoviruses in human fungal pathogens elicit an opposite effect, causing hypervirulence in cell culture and mouse infection models (Lau et al. 2018; Applen Clancey et al. 2020; Park et al. 2020; Rocha et al. 2025). Despite the emergence of mycoviruses as modulators of fungal pathogenesis, insights into how fungi control replication of these endosymbionts remain limited.

The S. cerevisiae L-A virus is the most comprehensively studied mycovirus, belonging to the broadly dispersed Totiviridae family of endogenous double stranded RNA (dsRNA) viruses. The 4.6 Kb L-A genome is encapsidated within a viral particle and extruded L-A transcripts encode for the capsid proteins (Gag) that form it. The L-A transcript also encodes a Gag-pol fusion protein that contains the canonical RNA dependent RNA polymerase (Pol) found in all RNA viruses. Each particle contains 1 or 2 Gag-pol proteins, which facilitates L-A genome replication and transcription. In many strains, L-A enables replication of satellite dsRNA segments called “M” that parasitize L-A viral particles. M satellites encode secreted toxins and a cell-autonomously acting toxin immunity factor. These “killer” toxins thus cause lethality in neighboring sensitive yeast (Wickner et al. 2013).

L-A replication is maintained at a low level through the SKI2, 3, and 8 genes, which encode subunits of a conserved translational surveillance complex that facilitates 3′-5′ exonucleolytic RNA decay and opposes the translation of transcripts that lack polyA tails like those encoded by L-A (Toh et al. 1978; Brown et al. 2000; Searfoss and Wickner 2000). Separate pathways of L-A attenuation act through Xrn1, a 5′-3′ exoribonuclease that degrades RNAs lacking 5′ methyl caps, and Nuc1, a versatile nuclease localized to the mitochondria (Ball et al. 1984; Rowley et al. 2016; Gao et al. 2019; Chau et al. 2023). Although L-A is asymptomatic in wild-type cells, in cells lacking NUC1, XRN1, and/or SKI genes, high levels of L-A and/or of Killer toxin cause lethality at extreme temperatures or in meiotic spore progeny (Ridley et al. 1984; Edwards et al. 2014; Gao et al. 2019; Chau et al. 2023).

Por1 is a well-known mitochondrial protein, but less well-characterized in the context of yeast antiviral responses. Earlier studies showed that L-A viral particles accumulate to high levels in por1Δ mutant cells grown on the nonfermentable carbon glycerol for several days (Dihanich et al. 1987, 1989). POR1 encodes the yeast voltage dependent anion channel (VDAC), a highly abundant mitochondrial membrane protein. VDAC beta-barrel proteins span the outer mitochondrial membrane and control small molecule flux. POR1 is essential for respiratory growth at high temperatures and plays important roles in mitochondrial osmotic stability, phospholipid metabolism, autophagy, mitochondrial protein import, and movement of phospholipids from one membrane bilayer leaflet to the other, a process known as lipid scrambling (Blachly-Dyson et al. 1997; Sanchez et al. 2001; Miyata et al. 2018; Sakaue et al. 2019; Broeskamp et al. 2021; Jahn et al. 2023; Takeda et al. 2025). In some contexts, VDACs assemble into higher order oligomers thought to facilitate movement of larger molecules, and recent findings show that Por1 lipid scramblase activity requires its oligomerized form (Jahn et al. 2023; Takeda et al. 2025). If and how any of these functions relate to L-A repression is unknown.

Here, we show that POR1 prevents L-A from accumulating to high levels in stationary phase cells following sustained culture in standard growth media. Using genetic and biochemical experiments, we elucidate that POR1 represses L-A by preventing hyperactivation of Snf1, the yeast homolog of AMP-activated kinase (AMPK). By investigating Snf1 signaling, we show that this POR1-SNF1 regulatory mechanism represses L-A replication in stationary phase cells by limiting amino acids. Our findings identify a novel POR1-SNF1 regulatory mechanism in budding yeast that may be similarly functional in other eukaryotes.

Materials and methods

Yeast strains, media, and plasmids

Standard S. cerevisiae genetic and strain manipulation techniques were used for strain construction. All strains are derivatives of BY4742 constructed through crossing and dissection. Generation of L-A cured strains was achieved using a previously described genetic backcrossing strategy (Gao et al. 2019; Chau et al. 2023). Refer to Supplementary Table 2 for all yeast strains used. For all experiments, strains were first grown in log phase for at least 10 doublings by serially splitting back of cultures. Time 0 for all stationary phase experiments corresponded to these log phase cells at an OD₆₀₀ of 1. All experiments were carried at 30 °C in synthetic complete media (SC; 0.348% yeast nitrogen base, 1% ammonium sulfate, and 2% glucose) with the appropriate amino acid powder mix (Sunrise Science) unless otherwise specified.

Measurement of yeast growth and physiological activity

Saturated cultures were diluted to an OD₆₀₀ of 0.1 in 200 µL in a 96-well plate. The plate was sealed a Breathe-Easy membrane (MilliporeSigma) and growth curve data was generated at 30 °C using an S&P growth curve robot (S&P Robotics Inc). Plates were shaken, and the optical density readings were taken every 15 min for 24 h. The data were plotted using R studio ggplot2. Glucose and ethanol concentration in media were measured using assay kits from MyBioSource (MBS8243232 and MBS8309715) according to the manufacturer's instructions. The luminescent signals were detected using Varioskan LUX Multimode Microplate Reader (ThermoFisher).

Protein extraction and western blotting

Cells were harvested at indicated time points and permeabilized with 0.1N NaOH at room temperature for 5 min. The cells were then pelleted and resuspended in SDS/PAGE buffer. Cells were disrupted by bead-beating for 3 min before heating at 100 °C for 10 min. The samples were centrifuged to isolate the soluble fraction for western blotting. Protein concentrations were determined with an RC/DC assay (BioRad 5000121). For Snf1-pT210 western blotting, cells were boiled at 100 °C for 3 min before extraction.

Equal amounts of protein were electrophoresed on 10% SDS-PAGE gels and transferred to polyvinylidene difluoride membranes. Membranes were incubated in primary antibody at 4 °C overnight and probed with 1:3,000 horseradish peroxidase (HRP)-conjugated horse anti-mouse (7075; Cell Signaling Technology) or goat anti-rabbit (7074; Cell Signaling Technology) secondary antibody. The proteins were detected with Luminata Forte Western HRP Substrate (EMD Millipore) and imaged with the Bio-Rad ChemiDoc XRS + system. Images were processed with the Image Lab software package (Bio-Rad). The primary antibodies and their dilutions were 1:1,000 anti-FLAG M2 (F1804; Sigma–Aldrich), 1:1,000 anti-VDAC1/Porin (ab110326; Abcam), 1:5,000 anti-Pgk1 (ab113687; Abcam), 1:2,000 anti-L-A Gag (obtained from Reed Wickner), and 1:1,000 anti-Phospho-AMPKα (Thr172) (2535; Cell Signaling Technology).

To quantify the relative fold change of protein levels, band intensities of protein of interest were measured in ImageJ and normalized to the band intensities of the corresponding loading control, Pgk1 protein. The average relative fold change and standard deviation was plotted using R studio ggplot2.

Reverse transcription-quantitative polymerase chain reaction

RNA was prepared from 10 ODs of cells harvested from 7-d cultures and used for reverse transcription-quantitative polymerase chain reaction (RT-qPCR) as previously described with some modifications. Briefly, harvested 7-d cell pellets were resuspended in Trizol (15596026; Invitrogen) and subjected to bead-beating (Mini Bead beater, Biospec Products) for 1 min followed by 2-min incubations on ice for 8 cycles. Samples were then incubated for 30 min at 65 °C in acidic phenol (P4682; Sigma–Aldrich), SDS, and buffer AE (10 mM Tris-HCl, 0.5 mM EDTA pH 9.0) solution. The phase-separated supernatant was washed with chloroform and precipitated overnight. The precipitate was then washed in 70% ethanol and dissolved in water. RNA samples were purified with the RNeasy Mini Kit (74104; Qiagen), and residual DNA was digested with DNase I (79254; Qiagen). Nine hundred nanograms of RNA were reverse transcribed using random nonamers and Maxima H Minus Reverse Transcriptase (EP0753; Thermo Fisher). The cDNA product was isolated by alkaline hydrolysis and treated with RNase A. Subsequently, qPCR was performed on 1/20 dilutions of cDNA product with the SensiFAST SYBR Hi-ROX Kit (BIO-92005; Meridian Bioscience) on the CFX384 platform (BioRad). The data were plotted using R studio ggplot2.

Spot analysis

Yeast strains were grown for 7 d at 30 °C to the stationary phase. Each strain was diluted to an OD₆₀₀ = 0.4, serial diluted 4 times by 10-fold and spotted onto agar plates containing synthetic complete media supplemented with 2% glucose with or without 0.1 mg/mL azetidine-2-carboxylic acid (AZC), as indicated.

Metabolite extractions and measurements by LC-MS/MS

Strains were grown overnight in YPAD. Saturated cultures were diluted to 0.02 and 0.05 OD for wild type and por1Δ, respectively, and 5 ODs of cells were collected at the 24-hour time point. Intracellular metabolites were extracted from yeast cells, and quantitatively estimated using targeted, liquid chromatography–tandem mass spectrometry (LC-MS/MS) approaches described earlier, with samples collected from the respective genotype and at the indicated time point (Walvekar et al. 2018; Rashida et al. 2021). Specifically, equal numbers of cells (∼2 × 10^7) were quenched in extraction buffer (60% methanol), extracted in 75% ethanol and dried down using a speed vacuum (rotatory evaporator). Metabolites were resuspended in mass spectrometry grade water and 10 µl sample was injected for LC-MS/MS and separated using Synergi 4-µm Fusion-RP 80 Å (150 × 4.6 mm) LC column (Phenomenex, 00F-4424-E0). Solvents used: 0.1% formic acid in water (Solvent A) and 0.1% formic acid in methanol (Solvent B). An AB Sciex QTRAP 5500 with Shimadzu Nexera series UPLC system was used. Mass spectrometry data were acquired using analyst 1.6.2 software (Sciex). For analysis, Multi-Quant version 3.0.1 and Peak View version 2.0 were used.

Results

POR1 represses Snf1/AMPK signaling in stationary phase to prevent L-A replication

Dihanich et al found that POR1 repressed L-A after cells had adapted to respiratory metabolism over several days (Dihanich et al. 1987). To confirm these findings in the reference S288c strain background that is naturally infected with L-A, we first characterized growth and metabolism of isogenic wild type and por1Δ strains through 7-day batch cultures. As expected, logarithmically growing wild-type cells vigorously fermented glucose to produce ethanol and ceased growth when glucose was exhausted (Fig. 1a and 1b). While por1Δ displayed slower logarithmic growth and an according reduced rate of glucose consumption and ethanol production, both wild type and por1Δ expended all the glucose by 20 h of culture (Fig. 1a and 1b). Over the course of the following 6 d, both wild type and por1Δ consumed all the ethanol indicating robust respiratory metabolism (Fig. 1b).

Fig. 1.

Por1 represses L-A replication in stationary phase cells. a) Growth curves and media glucose concentration of wild type and por1Δ strains is shown. Cells were grown in YPAD at 30 °C for 24 h. Density measurements of cell cultures were taken every 15 min, and glucose concentration in the media was measured at indicated time points. n = 3. b) Media ethanol concentration of these exact cultures was measured. c) Western blotting of L-A gag and Pgk1 protein levels in the indicated strains. Samples were collected from cultures grown at the indicated time points in SC media. Molecular weight markers are indicated on the right. d) Quantification of L-A Gag levels normalized to Pgk1 protein level in indicated strains cultured for 7 d. Relative fold change to wild type and standard deviation are shown. n = 8. ** P < 0.01 using unpaired student's t-test. e) RT-qPCR quantification of L-A RNA normalized to endogenous ACT1 RNA in the indicated strains cultured for 7 d. Mean RNA level and standard deviation are shown. n = 3. * P < 0.05 using unpaired student's t-test.

To investigate POR1 repression of L-A, we measured L-A Gag levels using western blots in wild type and por1Δ strains at intervals using these same culture conditions. We found a large increase of Gag levels in por1Δ cells by 7 d of incubation (Fig. 1c). In many replicates shown below, wild-type Gag levels were undetectable in comparison to the large amount observed in por1Δ. We measured relative Gag levels in 8 replicates in which Gag was detectable in the wild-type strain and found a 20-fold increase in por1Δ, though this seems likely to be an underestimate (Sure 1D). To confirm that this increase in Gag protein reflects higher viral copy number, we measured L-A RNA levels using RT-qPCR, which revealed a 60-fold increase in por1Δ compared with wild type (Fig. 1e). S288c is infected with another totivirus called L-BC that is maintained at a significantly reduced copy number compared with L-A (Sommer and Wickner 1982). RT-qPCR quantification of L-BC did not reveal any differences, perhaps suggesting that L-A outcompetes L-BC (Supplementary Fig. 1). These findings confirm that Por1 represses L-A viral replication in respiratory cells that have exhausted nutrients. These are typically referred to as stationary phase cells, known for their metabolic quiescence and enhanced stress resistance (Werner-Washburne et al. 1993).

Progression into stationary phase involves massive gene expression reprogramming characterized by repression of nearly all transcription but activation of catabolic genes (Radonjic et al. 2005). The PAS family kinase Rim15 and the AMP-activated protein kinase Snf1 govern much of this gene expression program through phosphorylation of downstream transcriptional regulators (Galdieri et al. 2010). To investigate the roles of SNF1 and RIM15 for viral replication in stationary phase, we compared the levels of L-A Gag in deletion mutants of these genes by themselves or in combination with por1Δ. Deletion of SNF1 completely reversed the Gag accumulation phenotype of por1Δ while rim15Δ had a comparatively minor effect (Fig. 2a). This epistatic relationship suggests that Por1 represses Snf1 and that derepressed Snf1 promotes L-A replication. To test this, we measured the accumulation of the active isoform of Snf1 phosphorylated on threonine-210 (Snf1-pT210) using western blotting experiments with an antibody that recognizes it (McCartney and Schmidt 2001). Supporting our model, we detected significantly higher levels of Snf1-pT210 throughout stationary phase in por1Δ cells compared with wild type (Fig. 2b and Supplementary Fig. 3a to c). Moreover, Por1 levels were strongly induced in stationary phase exactly when our genetic results show that it represses Snf1-pT210 (Fig. 2b and Supplementary Fig. 3b). These findings demonstrate that Por1 represses Snf1 activation in postdiauxic stationary phase cells and that SNF1 function is required for L-A hyperactivation.

Fig. 2.

Por1 negatively regulates Snf1 activity to control L-A replication. a) Western blotting of L-A Gag and Pgk1 protein levels in the indicated strains. Samples were collected from 7-d cultures grown in SC media. Molecular weight markers are indicated on the right. b) Western blotting of phosphorylated Snf1, Snf1-FLAG, Pgk1, and Por1 protein levels in the indicated strains. Samples were collected from cultures grown for the indicated time points in YPAD media. Molecular weight markers are indicated on the right. c) Schematic of Snf1 pathway. The Snf1 complex contains the alpha subunit Snf1, one of the 3 beta subunits Sip1/2 and Gal83, and the gamma subunit Snf4. The Snf1 complex regulates expression of target genes Pck1, Fbp1, and Icl1 by phosphorylating downstream transcription factors Cat8 and Sip4. d) Western blotting of L-A Gag and Pgk1 protein levels in the indicated strains. Samples were collected from 7-d cultures grown in SC media. Molecular weight markers are indicated on the right. e) Western blotting of L-A Gag and Pgk1 protein levels in the indicated strains. Samples were collected from 7-d cultures grown in SC media. Molecular weight markers are indicated on the right.

In addition to its activating phosphorylation, Snf1 is controlled through its association with cofactors in a conserved heterotrimeric complex (Fig. 2c). The gamma subunit Snf4 binds to Snf1's autoinhibitory domain to relieve Snf1's autoinhibition, while the beta subunits, Sip1, Sip2, and Gal83, control Snf1 activity in different subcellular localizations (Hedbacker and Carlson 2008). We assessed the levels of L-A Gag in por1Δ strains lacking these proteins to discern the contribution of Snf1 subunits for viral control. Like with snf1Δ, snf4Δ completely reversed the accumulation of high Gag levels caused by por1Δ (Fig. 2d). While none of the beta subunit deletions similarly reversed the por1Δ phenotype, gal83Δ caused a visibly decreased abundance of Gag (Fig. 2d). Gal83 controls Snf1 function in the nucleus (Vincent et al. 2001). These findings show that Snf1 cofactors are required for its proviral function when it is hyperactivated in cells lacking POR1.

Snf1 promotes L-A replication through its glyoxylate cycle targets

To elucidate how hyperactivated Snf1 promotes L-A replication, we investigated its well-characterized targets, the transcription factors Adr1, Cat8, and Sip4, which drive the expression of genes involved in alternative carbon utilization in glucose starved cells (Fig. 2c) (Hedbacker and Carlson 2008). To determine if these transcription factors mediate Snf1's proviral activity, we assessed L-A Gag levels in double mutants combining por1Δ with cat8Δ, sip4Δ, or adr1Δ. While adr1Δ had no consequence, the high Gag levels of por1Δ cells were reduced when CAT8 or SIP4 were deleted with a marked effect caused by cat8Δ (Fig. 2e). Cat8 and Sip4 regulate overlapping sets of genes involved in gluconeogenesis and the glyoxylate cycle (Fig. 2c) (Hiesinger et al. 2001; Roth et al. 2004). To test if Cat8 and Sip4 redundantly mediate the proviral function of Snf1, we combined deletions of their shared targets FBP1, PCK1, and ICL1, with por1Δ. FBP1 and PCK1 encode gluconeogenic proteins while ICL1 encodes a key protein of the glyoxylate cycle. We found that the high Gag levels of por1Δ cells were completely reverted by icl1Δ with only weak or no consequence of fbp1Δ or pck1Δ (Fig. 3a). Collectively, these findings show that POR1 prevents hyperactivation of Snf1 kinase in stationary phase cells, and that activated Snf1 promotes L-A replication in a manner dependent on its downstream target gene ICL1.

Fig. 3.

L-A replication in por1Δ stationary phase requires the glyoxylate cycle. a) Western blotting of L-A Gag and Pgk1 protein levels in the indicated strains. Samples were collected from 7-d cultures grown in SC media. Molecular weight markers are indicated on the right. b) Schematic of the glyoxylate and TCA cycles. The intermediates and flow of carbon through both cycles are indicated. Enzymes tested to be required for viral replication are shown in red, while enzymes dispensable for viral replication are shown in blue. c) Heat map of normalized fold change amino acid levels measured with LC-MS/MS (n = 3). Prototrophic strains of the indicated genotype were used for the metabolic studies. Samples were collected from 24-hour cultures grown in SC media.

Icl1 catalyzes the conversion of isocitrate to succinate and glyoxylate, a key intermediate of a TCA-cycle shunt known as the glyoxylate cycle found in bacteria, fungi, plants, and some invertebrates (Fig. 3b) (Dolan and Welch 2018; Chew et al. 2019). The requirement of ICL1 for L-A hyper-replication thus implicates the glyoxylate cycle in this process. We tested this through double mutant analysis combining por1Δ with other glyoxylate cycle gene deletions. As with icl1Δ, we found that mls1Δ, mdh2Δ, or cit2Δ reversed the por1Δ hyper-L-A phenotype (Fig. 3b and Supplementary Fig. 3a; Supplementary Table 1). While Icl1, Mls1, and Cit2 all act in the cytosol, glyoxylate metabolism also occurs within peroxisomes through Dal7 and Mdh3 as well as the cytoplasmic/peroxisomal dual-localized Mdh2 (Kunze et al. 2006; Gabay-Maskit et al. 2020). Deletion of MDH2, but not of DAL7 or MDH3, reversed L-A hyper-replication caused by por1Δ (Fig. 3b and Supplementary Fig. 3a; Supplementary Table 1). MDH2 and MDH3 encode malate dehydrogenases, and a third, Mdh1, localizes to the mitochondria where it functions in the TCA cycle. Deletion of mdh1Δ similarly failed to revert the por1Δ phenotype (Fig. 3b and Supplementary Fig. 3b; Supplementary Table 1). These findings show that genes encoding cytosolic glyoxylate cycle proteins are crucial for L-A hyper-replication in por1Δ cells.

Succinate produced by the glyoxylate cycle can be funneled to mitochondrial TCA-cycle metabolism (Fig. 3b). As the glyoxylate and TCA cycles are thus intertwined, we further tested the role of TCA-cycle genes for L-A Gag accumulation caused by por1Δ. Like with mdh1Δ, deletion of most TCA-cycle genes failed to revert the por1Δ phenotype (Fig. 3b and Supplementary Fig. 3b, 3c; Supplementary Table 1). However, the TCA-cycle mutants fum1Δ and sdh1Δ did revert por1Δ (Fig. 3b and Supplementary Fig. 3b, 3c; Supplementary Table 1). As Fum1 and Sdh1 act early in the TCA cycle to metabolize succinate, a plausible explanation for these findings may be through succinate buildup causing inhibition of Icl1 and/or other glyoxylate cycle enzymes (Honer Zu Bentrup et al. 1999). These findings refine a model in which Por1 repression of Snf1 prevents glyoxylate cycle activation that promotes L-A replication.

The POR1-SNF1 regulatory system prevents L-A hyper-replication by limiting amino acid availability

How might the glyoxylate cycle promote L-A replication in por1Δ cells? The glyoxylate cycle enables utilization of 2-carbon units during glucose starvation, producing metabolic intermediates that support gluconeogenesis and amino acid biosynthesis (Chew et al. 2019). An increased glyoxylate cycle will lead to more oxaloacetate, which can directly be converted to aspartic acid (which in turn supports the biosynthesis of multiple amino acids, as well as sustaining alpha-ketoglutarate production leading to glutamate/glutamine synthesis). We therefore hypothesized that enhanced amino acid availability caused by increased glyoxylate cycle flux may fuel L-A replication in por1Δ stationary phase cells. To test this, we first used LC-MS/MS to compare the levels of amino acids in postdiauxic wild type and por1Δ strains 24 h postinoculation. Consistent with our hypothesis, we observed substantially increased steady-state pools of multiple amino acids in por1Δ (Fig. 3c).

We further investigated amino acid synthesis enzymes that utilize glyoxylate cycle produced precursors. Glyoxylate can be converted to glycine by the alanine-glyoxylate aminotransferase Agx1 while oxaloacetate is converted to aspartate through the mitochondrial and cytosolic localized aspartate aminotransferases Aat1 and Aat2, respectively (Fig. 4a) (Morin et al. 1992; Verleur et al. 1997; Schlosser et al. 2004). In contrast to glycine, aspartate is a versatile amino acid that can be used to produce numerous additional amino acids (Fig. 4a). To test if glyoxylate cycle intermediates promote amino acid synthesis that enhance L-A replication, we combined por1Δ with deletions of AAT1, AAT2, and AGX1 to test for reversion of the por1Δ phenotype. While deletion of AAT1 or AGX1 had little effect on the levels of L-A Gag, aat2Δ reverted the accumulation of Gag in por1Δ (Fig. 4b). Like the required glyoxylate cycle proteins, Aat2 is cytosolic, suggesting that L-A replication in stationary phase requires the synthesis of aspartate from oxaloacetate in the cytosol.

Fig. 4.

Amino acid synthesis produces the biomolecules necessary for viral replication in a stationary phase. a) Schematic of amino acids synthesis pathways from the glyoxylate cycle. Enzymes tested to be required for viral replication are shown in red, while enzymes dispensable for viral replication are shown in blue. b–d) Western blotting of L-A Gag and Pgk1 protein levels in the indicated strains. Samples were collected from 7-d cultures grown in SC media. Molecular weight markers are indicated on the right. Additionally, samples were either supplemented with the indicated amount of (c) additional aspartic acid or (d) amino acids. d) The ponceau stain of the blot is included below. The protein ladder and corresponding molecular weight are indicated on the left.

If Aat2 promotes L-A replication in stationary phase by promoting the synthesis of aspartate, then aspartate supplementation in the media might restore high L-A viral load in por1Δ aat2Δ. Confirming this prediction, por1Δ aat2Δ cultured in increasing amounts of aspartate showed a dose responsive increase in L-A Gag levels (Fig. 4c). Since aspartate supplementation was insufficient to fully restore Gag levels in aat2Δ por1Δ, we tested if supplementing with the full complement of amino acids further enabled L-A replication. Remarkably, we found that aat2Δ por1Δ supplemented with 10× amino acids restored Gag accumulation comparable to as in a por1Δ single mutant (Fig. 4d). By treating these western blot membranes with the generic protein dye ponceau S, we found that Gag is the only protein that accumulates to high levels in these cells (Fig. 4d).

High L-A levels are known to cause proteostatic stress that sensitizes cells to the amino acid analog AZC (Chau et al. 2023). To test if high L-A has a similar consequence in stationary phase cells, we assayed them for growth following spotting onto SC plates with or without AZC. Wild type and por1Δ strains exhibited moderate and severe growth defects in the presence of AZC respectively and the severe growth defect of por1Δ was reversed in strains cured of L-A (Supplementary Fig. 4). These results suggest that high viral burden in por1Δ cells cause proteostatic stress.

To summarize, here we identify a new yeast antiviral mechanism controlled through Por1/VDAC mediated inhibition of Snf1/AMP Kinase in stationary phase cells. Consistent with many studies showing that Snf1 is activated by glucose starvation, we only observe accumulation of the activated phosphorylated form of Snf1 in postdiauxic phase cells that have consumed all the glucose. Following this transition to glucose starvation, por1Δ cells exhibit dramatically increased levels of Snf1-pT210 as well as a 60-fold increase in L-A, suggesting that activated Snf1 promotes L-A replication. Supporting this hypothesis, genetic epistasis experiments show that Snf1 acts with its cofactors and Cat8/Sip4 transcription factor targets to elevate L-A replication in por1Δ cells. Through LC-MS/MS, supplementation and genetic experiments, we identify the glyoxylate cycle and amino acid production as the essential outputs of Snf1 signaling that foster L-A replication (Fig. 5).

Fig. 5.

Model of Por1 repression on Snf1 pathway to limit L-A viral replication. In nutritionally starved cells, the AMPK Snf1 is activated to induce the glyoxylate cycle for the synthesis of amino acids. In wild-type cells, Por1 negatively represses Snf1 activity to prevent overproduction of amino acids in stationary phase cells. However, when this process is not properly regulated in por1Δ strains, the L-A virus hijacks the excess amino acid produced to enhance its replication success.

Discussion

One of the distinguishing features of mycoviruses is their endemic presence in strains once an infection is established. These infections persist on evolutionary timescales and can adaptively shape fungal physiology raising questions about how these long-term host-virus interactions are mediated in fungi (Wu and Li 2025). Here, we identify POR1-SNF1 signaling as a new yeast mechanism that prevents hyper-replication of the L-A totivirus through its control of cellular metabolism. Of note, we only observe POR1-SNF1 signaling in nutrient exhausted stationary phase cells that have exited the cell cycle. In nature, fungal cells likely encounter such conditions frequently, representing a point of vulnerability where mycoviruses could opportunistically replicate in their dormant hosts. Physiological control mechanisms that limit biomolecules essential for viral replication may thus represent a broadly useful strategy. Indeed, amino acid restriction acts as an antiviral strategy for bacterial phage and human HIV infections (Jiang et al. 2017; Fitzpatrick et al. 2025). Given that Snf1 controls allocations toward amino acids after glucose is depleted, any mechanism that can modulate this might provide advantages to the virus during infections (Rashida et al. 2021). Despite these connections however, our studies do not necessarily indicate that the sole function of POR1-SNF1 signaling is antiviral and it is easy to envision other potential roles of this metabolic control system.

A pressing question concerns the molecular mechanisms by which Por1 represses Snf1. Two studies showed Snf1 mitochondrial localization and physical association of Snf1 with Por1 in cells undergoing acute glucose withdrawal (Strogolova et al. 2012; Yi et al. 2017). Although neither study addressed Snf1 activation and signaling such as what we show here, they nevertheless are consistent with a proximal regulatory role of Por1 for Snf1.

Recent findings support a more vivid hypothesis of POR1-SNF1 signaling and point to Por1's control of lipid transport/metabolism as the underlying mechanism by which it represses Snf1. Thin layer chromatography and shotgun lipidomics experiments showed that deletion of POR1 caused myriad defects in bulk phospholipid levels, including reductions in phosphatidylethanolamine (PE) and cardiolipin along with increased phosphatidic acid (PA) and several others (Miyata et al. 2018; Broeskamp et al. 2021; Davis et al. 2023). Por1 controls these processes in part through its physical interactions with the Ups1/2-Mdm35 intramitochondria shuttles that control transport of PA and phosphatidylserine from the outer mitochondrial membrane to the inner membrane (Miyata et al. 2018). Notably, these studies investigated dividing cells, and our findings show Por1 function in stationary phase, a context known to involve massive lipid dynamics (Homann et al. 1987; Klose et al. 2012; Reinhard et al. 2023). A key insight comes from a study of Ups2-Mdm35 in stationary phase cells showing that its mutation caused PE accumulation and was associated with a modest increase in Snf1-pT210 levels, though not to the degree caused by por1Δ that we report here (Miyata et al. 2022). It was further shown that Snf1 binds to several phospholipids including PA and PE in vitro with the implied hypothesis being that Snf1 activity is controlled through sensing of lipids differentially controlled by Ups2-Mdm35 (Miyata et al. 2022).

All the above-mentioned studies preceded the discovery that Por1 possesses potent lipid scramblase activity in its oligomerized form and how POR1 influences lipid dynamics in stationary phase remains unknown (Jahn et al. 2023; Takeda et al. 2025). We report here that Por1 levels are strongly increased in stationary phase, which may promote the formation of Por1 oligomers that activate its lipid scramblase activity. Pharmacological inhibition of human VDAC has been shown to cause increased Snf1 activity and the human Snf1 homolog AMPK is known to be allosterically activated by long-chain fatty acyl-CoA esters (Head et al. 2015; Pinkosky et al. 2020; Ko et al. 2024). Evidence from both yeast and human thus suggest an underlying universality of POR1-SNF1 signaling.

Data availability

Strains are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article figures and tables.

Supplemental material available at GENETICS online.

Funding

This study was supported by grants from the Canadian Institutes of Health Research (PJT-18592) and Natural Sciences and Engineering Research Council of Canada (RGPIN-201) to M.D.M. S.L. is supported through a DBT-Wellcome Trust India Alliance fellowship (IA/S/21/2/505922).